Previous Page Table of Contents Next Page


The elusive tariff equivalent for the EU banana market


Pedro Arias, David Hallam, Lionel Hubbard and Pascal Liu[58]

This paper explores the effects of substituting the current EU tariff rate quota banana import system by a tariff-only system in 2006. The concept of "equivalence" is discussed, graphical analysis is used to explore the theory, and a dynamic, non-spatial partial equilibrium model of the world banana economy is built to test various policy scenarios. The paper argues that the various policy objectives pursued under the current import system could not be simultaneously achieved under a tariff-only system, but would require additional policy instruments. In particular, it claims that if quota C is binding, and elements exist to support this claim, then a tariff-only system that would maintain aggregate EU imports unchanged would result in a partial substitution of MFN banana imports by ACP.

1. INTRODUCTION

The European Union (EU) agreed with Ecuador and the United States in April 2001 to change its banana import regime from the current tariff rate quota (TRQ) to a tariff-only system no later than 1 January 2006. The potential effects of these changes on the world banana market and on the interests of the different groups of exporters and importers have led to extensive political and economic debate and controversy. Much of the controversy stems from the multiplicity of results obtained by analysts who tried to determine the tariff level that would replace the current system. The results are different not only because of the prices and methodological approaches used (price-gap versus modelling), but also because they differ on the policy objective (if any) that the regime should achieve. This paper aims at contributing to the debate by exploring the different meanings given to the term "equivalence" both in theoretical and empirical analyses. It also explores, with the help of an econometric model developed by FAO, the effect on countries benefiting from preferential trade agreements with the EU of substituting the current TRQ import regime by a tariff only.

The EU banana import system consists of a complex system of duty free and import quotas administered by import licenses. Bananas from ACP countries[59] can be imported duty-free under any quota. Bananas from other countries can only be imported under quotas A, B and the AQ, and must pay a tariff of €75 per tonne. No quantitative restrictions exist beyond quotas A, B, C and AQ, but imports are subject to a tariff of €680 per tonne for dollar banana imports, while ACP bananas benefit from a tariff preference of €300 per tonne. Although ACP bananas can enter the EU duty-free through any quota, they mostly do so under quota C because they usually cannot out-compete Latin American bananas in the other quotas.

Negotiations towards a tariff-only system

The EC notified the WTO of its intention to modify the MFN tariff for bananas in July 2004 and in October 2004 it suggested a single tariff of 230 euro per tonne for banana imports from most favoured nations (MFN), indicating that it had calculated this tariff level by computing the gap between internal and external EC prices. Formal negotiations with its main MFN suppliers under the auspices of the WTO started in November 2004. As no agreement was found, the EC notified the WTO of its proposed new tariff in January 2005.

The text of the waiver granted at the WTO Ministerial Conference in Doha in 2001 states that third country suppliers (e.g. Latin American suppliers) can call for arbitration should they disagree with the tariff proposal. On 30 March 2005, Ecuador, Colombia, Costa Rica, Panama, Honduras and Guatemala requested arbitration at the WTO on the level of tariff proposed. The complainants were later joined by Nicaragua, Venezuela and Brazil. In their award issued on 1 August 2005, the WTO arbitrators determined that the EC’s envisaged rebinding "would not result in at least maintaining total market access for MFN banana suppliers, taking into account all EC WTO market-access commitments relating to bananas". The European Commission then proposed a lower tariff of 187 euro per tonne and a duty-free quota of 775,000 tonnes for ACP bananas. This revised proposal was rejected by Latin American suppliers. On 26 September 2005, the EC requested a second WTO arbitration to determine whether its new proposal complied with the terms of the Doha Waiver. The arbitration ruled against the proposal. Following this ruling, the European Commission engaged again in negotiations with the above countries but no agreement could be found. On 25 November 2005, the Committee of Permanent Representatives of the EC approved a proposal of the European Commission for a tariff-only system from 1 January 2006 with an unbound tariff of 176 €/T and a duty-free quota of 775,000 MT for ACP bananas.

FAO 2005. Trade Policy Brief: Bananas

Virtually all ACP banana exports are destined for the EU. In addition, 18 percent of all Ecuadorean banana exports, 33 percent of those from Costa Rica, and 46 percent of those from Colombia are shipped to the EU. Latin American suppliers fear that, although their exports will no longer be constrained by quotas, an increased EU import tariff from the current €75 per tonne will erode their competitiveness vis-à-vis ACP suppliers, in particular African countries, and result in a loss of EU market share in the medium and long term. ACP countries, whose bananas enjoy duty-free access to the EU market under the Cotonou Agreement, are concerned that the price they receive following a regime change will not allow them to remain in banana production. Finally, EU producers who have 20 percent of the market, fear a fall in EU domestic price following the change in the import system would require increases in deficiency payments to levels unacceptable for both the WTO and the EU. This article concludes that it is unlikely that one single policy instrument (tariff only) will preserve the interests of all the major stakeholders.

2. ON THE EQUIVALENCE OF TARIFFS AND QUOTAS

Tariffs and quotas are policy instruments that can restrict imports into a particular market, and the possibility that they may be "equivalent" has been extensively studied. Bhagwati’s seminal paper (1965) on the non equivalence of tariffs and quotas under imperfectly competitive markets defines equivalence as follows: the quota gives rise to an implicit tariff rate which, if alternatively set as a tariff, generates the same level of imports as the quota. His first paper of 1965 proved that equivalence breaks down when domestic supply and/or import license holdings are monopolistic, and a later paper (1968) extended the proof that equivalence also breaks down when import supply is monopolistic. Incidentally, his later contribution was a reply to Shibata (1968) who proved that while equivalence in Bhagwati’s sense may not hold under monopolistic import supply, it may hold when equivalence is defined in terms of producing identical protective effects. Shibata’s paper raised awareness amongst policy makers on the fact that "equivalence" on Bhagwati’s term may not be adequate for the purpose of determining whether tariffs and quotas lead to equivalent policy effects. Bhagwati is fully aware of this problem and concludes that equivalence can only be resolved against its definitions, and that "this depends on the kinds of questions on which light is sought by analysis deploying the chosen definition".

This paper explores whether or not a tariff only import system exists that maintains the status quo of the current EU banana market. We define the status quo as the current trade flows in this market. The policy question (or equivalence) is formulated as follows: Does a tariff applicable only to MFN nations exist that reproduces the status quo? This definition is weaker than Bhagwati’s original equivalence, which was based on implicit and explicit tariffs. It is perhaps more closely related to "maintaining the level of protection" Shibata referred to in his 1968 paper, or to "maintaining the conditions of access" of the different traders the WTO arbitrators referred to in their Award (WTO 2005). Bhagwati (1968) had already formulated this question in saying that we would be "investigating whether the ‘real’ equilibrium corresponding to a tariff could be reproduced exactly by some quota alternatively imposed". The term "alternative" refers to methods other than those finding an implicit tariff such as price-gap computations. This paper concludes that elements exist to doubt that the status quo (equilibrium) can be reproduced with a tariff only system. The analysis focuses on the impact that tariff only would have specifically on ACP suppliers to the EU, who currently enjoy duty free-access and which are assumed to continue benefiting from this preference in the future. It will show that no tariff level simultaneously maintains the volume of imports, and the aggregate exports of ACP and of MFN suppliers.

3. MODELLING EU BANANA TRADE

For analytical purposes, the EU needs to be considered a large country in the sense that changes in its import policies are expected to have a significant impact on world trade. This section uses graphics to illustrate the broad effects of alternative EU import policies, namely the current TRQ and the future tariff-only regime. The demand schedule of the EU is an import demand, derived as the difference between total EU demand and EU domestic production. The following exposition assumes that EU domestic banana production is constant regardless of the import system, and therefore that the demand schedule of the EU is supplied solely by ACP and dollar banana suppliers. The analysis has as a point of departure free trade, and builds into the complexity of the current EU import system by explaining the workings of tariff only systems, tariff rate quotas, and tariff rate quotas with the possibility that a group of suppliers can enter the market duty free. The exposition assumes that both domestic and foreign markets are perfectly competitive,[60] and that exchange rates are fixed.

Tariff-only

Assuming that the point of departure is free trade, the imposition of an import tariff by a large country causes a reduction in its imports and an increase in its domestic price relative to free trade. Figures 1 and 2 illustrate the effects of the EU’s single tariff relative to free trade in the EU market. Figure 1 shows the free market situation. As stated before, import demand is derived as total demand minus domestic production, whereby the latter is assumed to be constant. Assume export supply to the EU market is divided between ACP and dollar sources. Figure 2 illustrates the effects of introducing a single tariff on dollar supplies while maintaining duty free access for ACP supplies. In this case, ACP suppliers would enjoy a level of protection equivalent to the magnitude of the tariff. Because of the tariff, total imports are lower than in the free market situation (decreasing from B to B1), and domestic prices increase (from Pf to P1). Higher prices in domestic markets, represented by P1, are to the benefit of domestic producers and those suppliers that have duty free access, but they are also to the disadvantage of consumers who face higher domestic prices relative to free trade. Prices received by dollar suppliers and their exports to the EU are also lower than under free trade (from Pf to P2 and from {B-Qf[ACP]} to B2 in Figure 2).

FIGURE 1
EU market behaviour under free trade


FIGURE 2
EU market behaviour following the imposition of an import tariff

Tariff rate quotas

Tariff rate quotas (TRQ), or multiple-tier tariffs, involve the specification of an import quota or quotas with different tiers of tariffs charged according to the volume of trade in relation to the quota level(s). In the most simple case a single quota might be specified with a lower in-quota tariff on imports up to the quota volume and a higher over-quota tariff on imports above the quota volume. Figure 3 illustrates the workings of a tariff rate quota import system in a large country, which includes an in-quota tariff and an out-of-quota tariff. Figure 3 represents just one case of the three possible market equilibria that can be obtained under TRQs. These possible scenarios of market equilibria depend on the size of the quota, the tariff levels, and the configuration of demand and supply. Define quota rents as the difference between supply costs, inclusive of any tariff, and prices received by suppliers (this definition of quota rent is valid throughout this paper):

(i) The quota is binding (fill rate is 100 percent) and therefore domestic and world prices are affected and quota rents are generated (represented in Figure 3).

(ii) The quota is not binding (fill rate is less than 100 percent) and therefore the imposition of the quota does not affect domestic or world markets (although the tariff charged within it will), and no quota rents are generated.

(iii) The quota is overfilled (fill rate is more than 100 percent) with imports beyond the quota attracting the higher tier tariff and quota rents are generated on imports under the quota.

FIGURE 3
Market behaviour under a tariff rate quota import system (binding quota)

Pf = Free market price
B = Import under free market
Q = Import quota
P1 = c.i.f. price
P2 = P1 + quota rent + in quota tariff
Ptt = P1 + out of quota tariff

The banana import market of the TRQ system of the EU is similar to that depicted in Figure 3. Nevertheless, the regime complicates Figure 3 even further, in that a quota exists for duty free imports of a group of suppliers. Specifically it consists of the following:

(a) A quota (C) of 750 000 tonnes is reserved for duty free imports from ACP countries.

(b) Two quotas (A and B) of a total 2 653 000 tonnes are reserved for bananas of any origin and subject to a tariff of €75 per tonne (zero for ACP countries).

(c) A new quota "AQ" (called "Additional Quantity" by the EU) was created following the accession to the EU of ten countries in May 2004. This quota is subject to the same conditions as quotas A and B, and was set at 300 000 tonnes for the period May/December 2004 and 460 000 tonnes for 2005.

(d) No quantitative restrictions exist beyond quotas A, B, C and AQ, but imports are subject to a tariff of €680 per tonne for dollar banana imports, while ACP bananas benefit from a tariff preference of €300 per tonne.

These quotas are administered through import licenses, mainly allocated to importers on the basis of their past trade in bananas with the EU. Approximately 17 per cent of quotas A, B and AQ, and 11 per cent of quota C, are reserved for non-traditional operators. Licenses are traded and therefore quota rents are being generated in the import system.

Let us start by depicting the hypothetical working of the EU market under two different tariff only rates (€75 and €680/t), and no import quota restrictions. This scenario is represented in Figure 4, that shows shifts of the supply curve for those two import tariff levels. Different segments of these supply curves will eventually determine the boundary of the supply curve under TRQ, as shown in Figure 5.

Consider the relationships that exists between a TRQ where quotas are not binding and a tariff only import system. Assume that a TRQ import system of the EU whereby MFN suppliers pay an in-quota tariff of €75/t, ACP have duty free access, and no quota is binding. Since quotas are not binding, lifting the quota restrictions and putting in place a tariff for MFN suppliers equal to €75/t would have no impact on trade flows. In addition, ACP suppliers would have no incentive to expand trade because domestic prices would remain unchanged. This equivalence between TRQs and tariff only holds true up to a market situation where quotas start being binding. Figure 5 depicts the upper limit of a scenario where the equivalence between tariffs and quotas exist, namely a situation where quotas are just binding and no quota rents are generated.

Figure 5 approaches the real actual representation of the EU banana import market, but it is not yet an accurate picture because there are indications that quotas A, B and C are binding. ACP exports to the EU in 2003 were higher than quota C. Industry sources inform that import licenses are traded within all quotas in 2003, including quota C. Within quota C, licenses initially allocated to Caribbean suppliers are sold to African suppliers and in addition, African suppliers may be selling within quotas A and B. Quotas A and B are binding because it is well known that their import licenses are traded. Therefore, Figure 5 needs to be redrawn accordingly.

FIGURE 4
EU banana import market under two tariff levels, and no quota restrictions

  • Free Trade
    S1 = ACP supply to the EU
    S2 = MFN supply to the EU
    S3 = Total supply to EU a = Equilibrium
    B = EU imports
    B’ = EU imports from MFN suppliers
    B" = EU imports from ACP suppliers
    Pf = World price

  • Effect of tariff only €75/t (no quota restrictions, ACP duty free access)
    S4 = Total supply to EU b = Equilibrium
    P1 = EU domestic price and ACP supplier price
    P1’ = MFN c.i.f. price
    B1 = Total EU imports
    B1’ = EU imports from MFN suppliers
    B1" = EU imports from ACP suppliers
  • Effect of tariff only €680 (no quota restrictions, ACP duty free access)
    S5 = Total supply to EU
    P2 = EU domestic price and ACP supplier price
    P2’ = MFN c.i.f. price
    B2 = Total EU imports
    B2’ = EU imports from MFN suppliers
    B2"= EU imports from ACP suppliers

Figure 6 may be a more realistic representation of the banana market under the current import system, whereby exports by both ACP and MFN suppliers are quota constrained. The cost of import licenses to ACP suppliers defines the kink of the curve (c.i.f. for ACP), and therefore the value of the quota rent captured within quota C. The value of this quota rent would be the difference between the domestic price (P1) and the c.i.f. (P1") price paid for ACP bananas.

FIGURE 5
EU banana import market assuming quotas are just binding (no quota rents are generated)

Note: S1 is ACP supply; S3 is total supply under free trade; S4 is MFN supply under €75/t; S5 is MFN supply under €680/t. Point "b" represents the market equilibrium that can be reproduced either by a quota on MFN supplies of B1 or a tariff of €75/t applicable to MFN suppliers and duty free access for ACP suppliers.

Quota C is reserved for ACP suppliers only, and licenses to import under quota C cannot be used to import within other quotas. Since the rents on quotas A, B and C depend on the difference in ACP and MFN marginal production costs and the existence of the €75/t MFN tariff, in general quota AB and C rents are expected to be different (see Figure 6). Unfortunately no data exists to confirm their different values, and therefore it is not possible to identify with precision the kink of ACP supply in Figure 6 (point "e"). What matters for the analysis is, however, whether quota rents are generated within quota C. The fact that these may be different from those generated in quotas A and B is irrelevant for the conclusion of this paper.

Suppose we substitute the import system in Figure 6 by a tariff only system that leaves total EU imports unchanged. Suppose also that ACP bananas continue to enter duty free. Recall that the objective of the original exercise is to find a tariff level that leaves the current trade flows unchanged, namely a tariff that shifts the aggregate supply function "S3" to the left and intercepts the EU import demand at point "b". The actual outcome of a tariff only that replaces the system depicted in Figure 6 can be observed in Figure 7. It shows that while ACP exports to the EU were constrained by quota C in Figure 6, under the new system they are no longer restricted and therefore expand. More specifically, ACP exports would expand by an additional X1 represented in Figure 7. Since total EU imports remain unchanged, X1 also represents the volume of MFN exports displaced by ACP supply. Therefore, Figure 7 shows that if ACP suppliers would enjoy unlimited duty free access under the new system, then no single tariff exists that would maintain current trade flows. The magnitude of displacement of MFN by ACP bananas depends on where the kink of the ACP supply curve is located.

FIGURE 6
EU banana market under the current system assuming all quotas are binding

B = total EU imports under free trade
B1 = Total EU imports under TRQ
B1’ = MFN exports to EU under TRQ
B1" = ACP exports to EU under TRQ
Pf = World price under free trade
P1 = EU domestic
P1’ = cif for MFN suppliers
P1" = cif for ACP suppliers
QR = Quota rents

In summary, if ACP suppliers continue to have duty free access in the new import system, any quota rent captured within this quota in the current system (regardless of which market player is currently capturing it) will be transformed into tariff preference under tariff only. As long as quota C is binding and quota rents are generated within this quota, the substitution of the current import system by a tariff only system that maintains aggregate EU imports unchanged in 2006 would result in a partial substitution of MFN bananas by ACP.

4. THE MODEL

World banana trade is represented by a dynamic, non spatial, partial equilibrium model similar to others used in recent analyses of this market (see for example Borrell and Bauer 2004 and Guyomard et al. 1999). Prices are endogenous and derived from market clearing conditions. It is calibrated for the three year period 2000/02 and simulates banana supply and demand until 2010. Supply and demand equations are double-log, including partial adjustment as appropriate. World price is assumed to be equal to c.i.f. in EU15, and is estimated from FAOSTAT as the unit value of imports of EU15 ($547/t for the period 2000-2002). The internal EU price of bananas imported within quotas A and B is this price plus the quota rent, plus the in-quota tariff of €75/t.

FIGURE 7
Substitution of tariff rate quota by tariff only under the assumption that quota C is binding

4.1 Model architecture

The model includes 19 importing countries and/or regions, and 16 suppliers. Parameters were estimated using E-VIEWS from FAO and World Bank time series data. Where actual import and export prices were not available, export and import unit values were used. In the few cases where satisfactory parameter estimates could not be obtained, parameter values were assumed on the evidence of past studies. Macroeconomic data is from the World Bank and its projection into the future comes from FAO/OECD modelling studies. The model is implemented in Excel and written with elementary arithmetic symbols and formulas, to maximise transferability.

Three sets of equations represent three different markets:

(i) the Asian market, represented by exports from the Philippines and imports from Asian countries;
(ii) the EU market represented, by the EU and its ACP and dollar banana supplying countries; and
(iii) the rest of the world (ROW).

One could assume that the Asian market is independent from ROW and the EU because the major exporter (the Philippines) sells most of its bananas in Japan and China, and not in Europe or the United States. By the same token, Latin American countries sell very low volumes of bananas to China and Japan. Therefore policy changes in EU or ROW may not affect the Asian market, and vice-versa. The rationale for including Asia obeys to the original purpose of this model, which was to do a medium term projection of the world banana trade. The model assumes that Near East imports both from Latin America and the Philippines, substituting from either source according to their relative prices (Armington elasticity of substitution).[61] EU domestic supply is excluded from the model and assumed to be constant due to deficiency payments and constant productivity.

Supply equations

SjASIA

= f {PASIA, SjASIA(-1), gj}

SjACP

= f (PROW, SjACP(-1), t, QRC, gj)

SjROW

= f {PROW, SjROW(-1), gj}

Demand equations

DiASIA

= f { PASIA (XRi, ti), DiASIA(-1),Yi, gi, Pop}

DEU

= f {PROW(XR€/$, QRAB, tEU),YEU, gEU, Pop}

DiROW

= f { PROW (XRi, ti), DiROW(-1),Yi, gi, Pop}

Market clearing conditions

S[SjASIA] = S[DiASIA]

solves for PASIA

S[SjROW] + S[SjACP] = S[DiROW] + DEU

solves for PROW

Where "j" denotes exporting countries; "i" importing countries; ASIA is the Asian market; EU is the European Union; "ROW" is rest of the world; "S" is supply; "D" is demand; "PASIA" and "PROW" are banana prices in each market; "XR" is the exchange rate; "QRAB" is the quota rent generated in quotas A and B; "QRC" is the quota rent generated in quota C; "tEU" is the MFN in-quota import tariff of the EU; "-1" are lagged values; "Y" is income per caput; "g" is a time trend; and "Pop" is population; The inclusion of "g" in supply equations captures growth in supply due to technical change, economies of scale or efficiencies, while in demand equations it captures changes in consumer preferences.

In summary, the model assumes that export supply depends on the price of bananas, supply in the previous year (partial adjustment) and a time trend that captures the rate of growth of banana exports not due to price effects. Import demand depends on banana prices, demand in the previous year, income per caput, a time trend that captures the rate of growth of banana imports not due to price effects, currency exchange rates, import tariffs and population.

The model represents two chronological phases: Phase I is the TRQ system, and Phase II the tariff-only system. Therefore Phase I extends from the base period until 2005 and Phase II includes the years 2006 to 2010. The equations for calibration and Phase II are as described above, but those of Phase I are slightly different. In Phase I, EU and ROW markets are independent from each other, and the following identities apply:

a. EU demand is assumed to be equal to the aggregate of all quotas, and the equation is solved for the endogenous variable PEU;

b. ACP supply is fixed at the quota value of 750 000 tonnes and the identity yields the endogenous variable PACP;

c. ROW supply satisfies ROW demand plus quotas A, B and AQ, and the identity yields the endogenous variable PROW.

Following the accession of 10 countries to the EU in May 2004, a quota AQ of 300 000 tonnes for 2004 and 460 000 tonnes for 2005 are added to the EU (demand from these countries is accordingly eliminated from ROW from May 2004 onwards).

Equilibrium conditions in Phase I:

S[SjASIA]

=

S[DiASIA]

DEU

=

3.4 million tonnes in 2003 (Quotas A, B and C)


=

3.7 million tonnes in 2004 (Quotas A, B, C and AQ)


=

3.86 million tonnes in 2005 (Quotas A, B, C and AQ)

SjACP

=

[Quota C]

SSjROW

=

S[DiROW] + [Quotas A, B & AQ]

4.2 Unresolved issues

The model can be used to explore the effect of different tariff levels on world banana trade and in particular on developing country exports. This section outlines some of the problems encountered in its construction. A more detailed discussion of the problems associated with the use of models on world banana trade can be found in FAO (2005).

Any model involves a trade-off between making a realistic representation of markets and, at the same time, allowing for analytical tractability. The current model of necessity follows other efforts at analysing banana trade policy in making a number of simplifications concerning the nature of the banana market. It assumes perfect competition; there is perfect price transmission; and bananas have no substitutes.

Estimation results for the various parameters of the model were not always satisfactory in statistical terms. Most importantly, it was not possible to obtain consistent estimates of price and income elasticities of demand for the EU as a whole. Satisfactory estimates could be obtained for individual EU countries, but not for the aggregate. Short run elasticities for individual countries varied significantly from a low of -0.01 in the United Kingdom to a maximum of -0.37 in Germany. These relatively low values are consistent with the view of market analysts that banana demand in the EU is relatively inelastic. For the United States, estimates were not statistically significant, although the values computed were within the range of inelastic responses. The decision was therefore taken to assume that EU and United States demand parameters were similar: specifically an own price elasticity of -0.28 and an income elasticity of 0.5.[62] These values used are in line with those used in other recent models of the banana market (see FAO 2005).

A similar problem was encountered in the estimation of parameters for the new accession countries. A statistically significant income elasticity of 1.57 was obtained for Hungary, but not for Poland. The value assumed for the aggregate of these countries was 1.4, which can be considered conservative in the light of the estimate for Hungary. The economies of these countries are expected to grow significantly following their accession to the EU, and hence the model predicts an expansion of imports in the medium term to rates which may not be realistic (about 7 percent per annum).

The time series data also produced some questionable parameters for ACP suppliers. First, the time trends computed for Côte d’Ivoire and Cameroon are not consistent with an a priori qualitative analysis of their export response. It is not the estimates of the trend coefficients themselves, but the fact that Cameroon’s is lower than that of Côte d’Ivoire when in the latter production has been affected by civil strife in recent years. Secondly, own price supply elasticities of Caribbean exporting countries appear to be higher than those of their African counterparts. However, this may simply reflect the preponderance of downward adjustments in supply over the estimation period. Supply response to rising prices may be different from supply response to falling prices.[63]

TABLE 1
Parameters of ACP supply equations of the model

ACP supply equations

Intercept

Own Price

Partial Adj.

Trend (%p/a)

Belize

19.14

0.86

0.21

5.3

Dominica

6.28

0.37

0.58

0.0

Dominican Rep

2.48

0.30

0.75

2.8

Grenada

1.27

0.26

0.70

0.0

S. Lucia

1.96

0.68

0.53

0.0

S. Vincent & Grenadines

1.60

0.68

0.53

2.4

Cameroon

18.84

0.16

0.68

1.3

Côte d’Ivoire

9.11

0.30

0.66

2.8

Supply equations are expected to emulate the actual behaviour of suppliers following changes in market conditions. Since the equations are different from each other, one market signal can have different effects on different suppliers. The model may find that aggregate supply from ACP may remain unchanged under two different policy scenarios, but close scrutiny reveals that different players within this group of countries have behaved differently. In this particular case, a policy that maintains aggregate ACP exports to the EU unchanged also means that African ACP may partly displace Caribbean ACP suppliers. In summary, a tariff level or a quota for ACP that maintains ACP aggregate exports to the EU unchanged also results in changes in the relative market share of the suppliers.

A sensitivity analysis was carried out to test the robustness of the results to changes in various parameters. Preliminary runs suggest that the import tariff in the EU in 2006, the trend coefficients of supply functions, and the €/$ exchange rate have significant effects on model results. Conversely, the model appears to be robust to changes in own price elasticity of the EU demand function, assuming that it stays within the range of relatively inelastic demand responses.

5. SEARCHING FOR AN EQUIVALENT TARIFF

5.1 Under tariff-only

The model finds that as long as ACP countries continue to have duty free access, no tariff-only can maintain the status quo. The model finds through the years that span from the baseline (2000-2002) and until 2005 that quota C is increasingly binding. This is due to the following: the model includes trend elements in both supply and demand equations which represent changes in productivity and consumer preferences over time. Since EU imports from ACP are restricted by a quota C, and this quota is assumed to be filled in the baseline, an increasing gap is generated with time between the higher domestic price that consumers are willing to pay (due to income and population growth) and the lower production cost of ACP suppliers (due to their increasing productivity). The mismatch through the years between productivity growth and higher consumer preference is equivalent to generating quota rents within quota C. Prices consumers would be willing to pay for the product increase through the years because supply is fixed by quotas, while offer prices decrease because of increases in productivity, and therefore all quotas are increasingly binding through the years.[64]

The calibration of the model can be used to illustrate the failure of the model to find a tariff that maintains trade flows unchanged. Calibration requires two assumptions: the first one is the pass-through of world prices to the domestic EU market, and the second the pass-through of domestic EU prices to ACP suppliers. Domestic EU prices consist of three elements: c.i.f. prices of MFN suppliers (dollar bananas), the in-quota tariff (€75/t) and the quota rent generated in quotas A and B (which we can approximate to the price of import licenses). Assume that in baseline period (2000/02), the average c.i.f. price of dollar bananas was €576/t, the in quota tariff €75/t, and the value of an import license €100/t. Then the domestic EU price in the baseline period would be €751/t, with an implicit tariff of €175/t.

Assume that quota C is not binding. In this case, c.i.f. for ACP suppliers is the EU domestic price, meaning that these suppliers need the tariff preference plus the quota rents generated in quotas A and B to stay in production. In this case, ACP supply equations would be calibrated with a high price, and therefore ACP suppliers would be high cost suppliers. Therefore, the baseline calibration can be reproduced with the two import systems, namely it would be possible to calibrate the model using both a tariff rate quota system or a tariff-only, for both would reproduce the status quo. The preference enjoyed by ACP suppliers under the TRQ would be identical to the level of protection provided by the tariff-only, so they would not have an incentive to export more or less produce. Equally so, MFN suppliers would not have an incentive to expand exports to the EU under tariff only because their c.i.f. prices would remain unchanged in both scenarios.[65]

However, consider a situation where ACP suppliers have production costs below the domestic EU price. Consider for example that ACP suppliers need not only the in-quota tariff preference, but also 50 per cent of the equivalent quota rent generated in quotas A and B to cover production costs. In this case, the remaining 50 percent of the quota rent would be captured by suppliers along the chain, and valued at the price of import licenses within this quota. Then, the explicit tariff that would maintain ACP level of protection would be {€75/t + €50/t = €125/t}. However, this value is lower than the one needed to maintain MFN country exports to the EU unchanged (€175/t). In other words, if the tariff-only is fixed at €175/t, then ACP exports would expand. Therefore, the architecture of the model in the baseline is such that no tariff only reproduces the tariff rate quota system. This is precisely the effect that the model is generating through the years for ACP suppliers, as the growth coefficients increase their productivity, and income and population growth increase demand.

Table 2 shows preliminary results from the model as the EU changes its import regime from a TRQ to a tariff only in 2006, including the volume of EU imports and exports from both MFN and ACP suppliers for different tariff levels (€75, €140 and €230/t). Three issues signal that no tariff equivalent maintains the status quo:

The latter result, as explained earlier, is because the implicit tariffs for different market players change over time, and thus the explicit tariffs required to leave their trade flows unchanged in one particular year. For example in ACP, the increase in productivity (lower production costs) faces an increase in demand (due to population and income growth). Their combined effect results in an increasingly binding quota C, and a scenario as depicted in Figure 7 where the substitution of the implicit tariff by an explicit tariff that leaves EU imports unchanged results in a substitution of MFN bananas by ACP.

TABLE 2
Percent change (2006 relative to 2005) in EU imports, ACP and Latin American exports to the EU under different tariff and quota rent scenarios

Quota C rent as percentage of quota AB rent (€120/t)

Tariff

75

140

230


EU Imports

11.5

7.2

2

100%

ACP Exports to EU

1

4.5

9.1


LAM exports to EU

14

7.9

0.3


EU Imports

11.6

7.3

2.1

50%

ACP Exports to EU

3.8

7.5

12.4


LAM exports to EU

13.5

7.2

-0.4


EU Imports

11.7

7.4

2.1

0%

ACP Exports to EU

7.1

11

16


LAM exports to EU

12.8

6.5

-1.2


FIGURE 8
EU imports and ACP and Latin American banana exports for different tariff levels assuming ACP suppliers capture 100 per cent of quota rent

Note that for every scenario of quota rent capture by ACP suppliers, the model finds a tariff that leaves the import share of aggregate ACP and MFN suppliers unchanged. However, these tariffs would result in an expansion of EU imports, and therefore do not maintain the status quo.

5.2 Restricting ACP imports by a quota

Define status quo under the new import system as maintaining current trade flows. This scenario implies two different policy objectives: to maintain aggregate ACP and aggregate MFN exports to the EU unchanged relative to the current system. As already suggested in the conclusion of FAO’s Technical Note Nr. 3, at least two different policy instruments would be needed to achieve these two different policy objectives. A quota restriction on ACP imports, in addition to the MFN tariff, would be required to maintain the aggregate flows of ACP and MFN suppliers unchanged in 2006 relative to 2005.

However, suppose that the objective of the new import system were to maintain the interest of each supplier in 2006, as opposed to keeping aggregate volumes unchanged. Then, the inclusion of a quota constraining aggregate ACP exports, and imposing to MFN imports a tariff level that maintains their aggregate exports to the EU unchanged, would result in changes of the relative trade flows of different players within each category of supplier. For example even within the group of ACP suppliers, low cost suppliers would tend to displace high cost suppliers (see simulation results in Annex 1). In addition, domestic EU banana prices under tariff only will be subject to the same dynamics of the open banana market, including the volatility and long term decline of prices that characterize this commodity. Therefore, the displacement of high cost suppliers by low cost is likely to continue and to exacerbate in the long run (see Annex).

6. CONCLUSIONS

The analysis and model results presented in this paper show that it may be impossible to devise a tariff-only system that would maintain the trade flows of the current TRQ system. If quota C is binding, a tariff-only system that would maintain aggregate EU imports unchanged would result in a partial substitution of MFN banana imports by ACP. Conversely, a tariff-only that leaves the market share of aggregate MFN and aggregate ACP imports unchanged would be accompanied by an expansion of EU imports and a fall in the EU domestic price.

It appears therefore that the EC cannot simultaneously maintain the interests of all major players of the banana trade in 2006 with the use of only one policy instrument, as in this case a tariff-only. Even in the event that it combines tariff-only with a quota restriction for ACP suppliers, differences in competitiveness of suppliers, combined with highly volatile and declining banana prices, will result in progressive changes in their relative market share. Several policy instruments would be needed to maintain the interests of major stakeholders. The number of instruments would be as large as the number of policy objectives the EC intends to meet in the new import system.

Policy makers applying a tariff only to MFN suppliers and restricting ACP imports by a quota will claim that only practical questions remain unanswered. However, it precisely is in resolving these practicalities that the shape of the system of preferences is to be determined in the future. Consider for example how import licenses would be allocated within an ACP quota. Under the current system the value of an ACP import license depends not only on the tariff preference but also on the quota rent generated. However, under the tariff only the value of the import license would equate that of the tariff preference only (quota rents disappear). If licenses are auctioned, high cost (mostly Caribbean) suppliers would be outbid by low cost (mostly African) suppliers. If, on the other hand, licenses are distributed according to historical reference, then high cost suppliers will also be affected, but less than under an auction system. Under tariff only, ACP suppliers would maintain their margins of preference relative to MFN suppliers (equal to the tariff), but they also face a progressive fall in the EU domestic price in line with the fall in world prices.[66]

REFERENCES

Bhagwati, J. 196). More on the Equivalence of Tariffs and Quotas, American Economic Review 58, pp. 142-146.

Borrell, B. & Bauer, M. 2004. EU Banana Drama: not over yet. CIE, Canberra and Sydney.

FAO. 2005. Bananas: Is there a tariff-only equivalent to the EU tariff rate quota regime? Insights from economic analysis, Trade Policy Technical Note 3. FAO, Rome.

FAO. 2005. Bananas: Implications of EU tariff reform for producers, Trade Policy Briefs. FAO, Rome.

Guyomard, H., Laroche, C. & Le Mouël, C. 1999. Impacts of the Common Market Organization for Bananas on European Union Markets, International Trade, and Welfare. Journal of Policy Modeling I21(5):619-631.

Herrmann, R. & Sexton, R. 2001. Market Conduct and the Economic Impacts of a Tariff Rate Quota Policy: The European Banana Case. In Moss, C., G.B. Rausser, A. Schmitz, T. Taylor and D. Zilberman (eds.), Agricultural Globalization, Trade and the Environment. Kluwer Academic Press, Dordrecht.

Shibata, H. 1968. A Note on the Equivalence of Tariffs and Quotas, American Economic Review 58, pp. 137-142.

Vanzetti, D., Fernandez de Córdoba, S. & Chau, V. 2004. Banana Split: How EU Policies Divide Global Producer. Paper presented for FAO Informal Expert Consultation on Banana Trade Policies, Rome, 28-29 October 2004.

WTO. 2005. European Communities - The ACP-EC Partnership Agreement - Recourse to Arbitration Pursuant to the Decision of 14 November 2001; WT/L/616; 1 August 2005 (05-3440). WTO, Geneva.

ANNEX

EU imports, banana prices, ACP exports and export revenues under tariff-only systems of €230/t and €187/t, with and without ACP quota restrictions. These scenarios are extracted from Box 1 above, and represent two proposals made by the European Commission. They are presented for the sole purpose of showing model results that support the claim that no single tariff applicable to MFN suppliers (while maintaining ACP duty free access) would preserve current trade flows.

TARIFF ONLY
MFN tariff 230€/t
No ACP quota restriction

TARIFF ONLY + QUOTA C
MFN tariff 187€/t
ACP quota of 775 000 t

1. EU imports (kt)



2005

2006

2010

2005

2006

2010

Total

3 860

3 940

4 338

3 860

4 026

4 434

New Acc. Countries

523

600

781

523

634

828

From L.Am.

3 110

3 097

3 237

3 110

3 251

3 665

From ACP

750

842

1 100

750

775

775





2. Banana Prices (USD/t)



2005

2006

2010

2005

2006

2010

World price

525

513

460

525

522

472

EU domestic price

760

789

736

760

736

696





3. ACP exports to the EU (tonnes)



2005

2006

2010

2005

2006

2010

Belize

56 821

75 482

97 168

56 821

61 353

55 838

Dominica

23 573

25 457

26 850

23 573

23 320

18 440

Dominican Republic

116 661

131 525

200 684

116 661

122 392

134 820

Grenada

665

699

740

665

657

538

S. Lucia

50 580

58 697

63 548

50 580

49 813

33 134

S. Vincent

34 323

39 831

43 123

34 323

33 803

22 484

Cameroon

236 038

250 530

299 163

236 038

241 294

248 430

Côte d’Ivoire

231 339

260 455

369 208

231 339

242 368

261 316





4. ACP Export Revenues (thousand USD)



2005

2006

2010

2005

2006

2010

Belize

43 164

59 519

71 471

43 164

45 787

38 872

Dominica

17 907

20 073

19 749

17 907

17 404

12 837

Dominican Republic

88 621

103 711

147 610

88 621

91 339

93 855

Grenada

505

551

544

505

490

374

S. Lucia

38 423

46 284

46 742

38 423

37 175

23 066

S. Vincent

26 073

31 408

31 719

26 073

25 226

15 652

Cameroon

179 304

197 548

220 045

179 304

180 074

172 945

Côte d’Ivoire

175 735

205 374

271 566

175 735

180 876

181 916

Total

569 732

664 468

809 447

569 732

578 371

539 517

Source: model output
Assumptions: Quota rents in A and B in the baseline = €120/t.
ACP suppliers capture 50% of quota rent
Exchange rates of 1.2 $/€ in 2006 through to 2010


[58] Pedro Arias and Pascal Liu are Commodity Specialists and David Hallam is Chief in the Raw Materials, Tropical and Horticultural Products Service, Commodities and Trade Division, FAO. Lionel Hubbard is Senior Lecturer in Agricultural Economics in the University of Newcastle Upon Tyne. The authors are grateful for useful comments provided by Giovanni Anania.
[59] Countries in sub-Saharan Africa, the Caribbean and the Pacific that are former colonies of some EU member states and have signed the Cotonou agreement of economic partnership with the EU.
[60] For an analysis of market structure in the EU banana market see Herrmann & Sexton 2001.
[61] The world banana trade has also been explored with a multi-commodity Armington specification model by Vanzetti et al. (2004).
[62] A sensitivity analysis revealed that the model is robust to changes in the values of these parameters, provided that they remain within the range of relatively inelastic demand responses.
[63] The asymmetry of supply response is discussed in more detail in FAO 2005.
[64] In addition, the AQ (additional quantity) quota of imports created following the accession of 10 new countries in May 2004 is lower than the baseline, i.e. lower than the volumes at which their demand functions were calibrated in the baseline. This means that bananas are transhipped from EU15 to the new accession countries to cover the deficiency, putting more pressure on the prices of bananas imported in other quotas. In fact, numerous press reports claim that there was a considerable shortage of import licences in 2005.
[65] Assuming perfect competition in both domestic and world markets and consumer preference is identical for ACP and MFN bananas.
[66] Under a tariff only system the only difference between domestic and world prices is the tariff level, and therefore domestic prices will fall in line with the secular downward trend of world banana prices.

Previous Page Top of Page Next Page